GB1603462A - Catalyst supports and catalysts prepared therefrom - Google Patents

Description

PATENT SPECIFICATION ( 11) 1 603 462

9 ( 21) Application No 29046/80 ( 22) Filed 23 March 1978 + ( 62) Divided out of No 1603461 ( 31) Convention Application No 781379 CV 4 ( 32) Filed 25 March 1977 C:> ( 31) Convention Application No 781393 ( 32) Filed 25 March 1977-in -4 ( 33) United States of America (US) ( 44) Complete Specification published 25 Nov 1981 ( 51) INT CL 3 COIF 7/02 ( 52) Index at acceptance CIA 519 N 4 PB 5 ( 54) CATALYST SUPPORTS AND CATALYSTS PREPARED THEREFROM ( 71) We, W R GRACE & CO, a Corporation organized and existing under the laws of the State of Connecticut, United States of America, of Grace Plaza, 1114 Avenue of the Americas, New York, New York 10036, United States of America, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed to be particularly 5 described in and by the following statement:-

This invention relates to catalysts, especially to catalysts for converting harmful components of the exhaust gases of an internal combustion engine to less harmful entities, and to catalyst supports therefor.

A common form of catalyst consists of a solid support and a catalytic material 10 carried on the support; the solid support is then usually particles of a porous material-e g alumina The particles are typically of a size equal to spherical particles 1 mm in diameter up to 15 mm in diameter The porous material-e g.

alumina-may need to be shaped into particle form-e g spheres, spheroids, pills, cylinders-and we will refer to the material before it is shaped as "support 15 material" and after it is shaped as "formed support" The novel alumina monohydrate of this invention and its admixture with amorphous alumina are in this sense a support material; the microspheroidal particles, whether made from the support material of this invention or from a similar material, are in this sense a formed support; and these microspheroids when loaded with a catalytic agent are a 20 catalyst.

The activity, efficiency, stability, and durability of a catalyst in a reaction depend upon the chemical, physical, and structural properties of the catalyst precursors, i e, the support material and the formed support particles, and the nature and distribution of the catalytic material on the formed support Minor 25 variations in these properties may produce substantial differences in the performance of the catalyst Desirably, the properties of the support material that enhance catalytic activity are retained by the formed support particles In general, the formed support and catalyst comprising small amounts of the catalytic material on the support have essentially the same physical and structural properties with 30 slight differences due to the effects of the thermal activation of the catalyst.

The internal porous structure of the catalyst particles and their presursors determines the extent and accessibility of surface area available for contact of the catalytic materials and the reactants Increased pore size results in greater diffusion rates for reactants and products in and out of the catalyst particles and this often 35 results in improved catalyst activity However, the extent to which pore size can be advantageously increased is limited As the pore size is increased, there is a decrease in the surface area where the reactions take place A good catalyst should have a balanced combination of high specific surface area, cumulative pore volume, and macroporosity High macroporosity means a pore size distribution 40 with a relatively high proportion of pores having a diameter greater than 1000 A.

Further, alumina and formed alumina with a low density and consequent low thermal inertia will produce a catalyst that will reach reaction temperatures sooner.

Catalyst support material is frequently a porous refractory inorganic oxide, 45 such as silica, alumina, magnesia, zirconia, titania, and combinations thereof.

2 1,603,462 2 Alumina is a particularly desirable support material since it inherently has a high degree of porosity and will maintain a comparatively high surface area over the temperature range normally encountered in many catalytic reactions However, when used under high temperature conditions for long periods of time, overheating of the alumina may cause sintering and change in the crystalline phase of the alumine which reduce catalytic activity, for example, due to loss of surface area available for catalysis Alumina is used as a catalyst support in the form of a finely divided powder or of macrosize particles formed from a powder.

Since the physical and chemical properties of alumina are highly dependent on the procedures followed in its preparation, many preparation processes have been 1 developed in attempts to optimize its properties for use as a catalyst support material Alumina is frequently precipitated by combining a water-soluble, acidic aluminum compound which may be an aluminum salt such as aluminum sulfate, aluminum nitrate, or aluminum chloride, and an alkali metal aluminate such as sodium or potassium aluminate However, the properties of the resultant I compositions after washing and drying have generally been deficient in one or more of the properties of high surface area, macroporosity, phase stability, and low density.

Typical methods of making alumina with some of the characteristics desired of a support material are disclosed in U S Patents 2,988,520, 3,520,654, 3, 032,514 and 2 3,124,418 U S Patent 3,864,461 is particularly interesting as it discloses the production of a crystalline alumina of low bulk density which is identified by its xray bands as identical to pseudoboehmite, in some cases in admixture with a small amount of bayerite.

In addition to retaining the surface area, porosity, and density characteristics 2 of the starting alumina material, a process for the formation of macrosize alumina particles should produce formed alumina with low shrinkage and high attrition resistance and crush strength Conventional low density supports are generally deficient in structural integrity Unless stabilized, an alumina particle will undergo considerable shrinkage of its geometric volume when exposed to high temperatures 1 during use Excessive shrinkage produces unoccupied channels in the catalyst bed through which reactants pass without contact with the catalyst.

High attrition resistance provides structural integrity and retention of activity under conditions of mechanical stress During transfer, loading into the reaction zone, and prolonged use, the catalyst particles are subjected to many collisions 3:

which result in loss of material from the outer layers Attrition of the catalytically active layer present in the outer volume of the particles affects catalytic performance and also results in a decrease of the volume of the material in the reaction zone Volume loss by shrinkage and/or attrition of the highly compacted, tightly held particles in a fixed catalyst bed tends to loosen them and allow for 4 ( increased motion and collisions during vibration Once a packed bed becomes loose, attrition tends to increase During storage, the catalyst is often packed in large tall containers awaiting loading In order to withstand the forces generated by the weight of the particles above them, the catalyst must exhibit high crush strength.

The size, size distribution, and shape of the particles affect both structural 4 f integrity and catalytic activity These properties determine the volume of catalyst than can be packed in a fixed bed, the pressure drop across the bed, and the outer surface area available for contact with the reactants Finely divided alumina may be pelletized, tabletized, molded or extruded into macrosize particles of the desired size and shape Typically, the macrosize particles are cylinders of diameter about 5 ( 1/32 to 1/4 inch and a length to a diameter ratio of about 1: 1 to 3:1 Other shapes include spheroidal, polylobal, figure-eight, clover leaf, dumbbell and the like.

Spheroids offer numerous advantages as a catalyst support over particles having angular shaped surfaces with salients or irregularities, such as extruded cylinders Spheroidally shaped particles permit a more uniform packing of the 5 ' catalyst bed, thereby reducing variations in the pressure drop through the bed and in turn reducing channelling which would result in a portion of the bed being bypassed Another advantage in using particles of this shape is that the spheroids exhibit no sharp edges which will attrit during processing, transfer, or use.

One of the most described methods for producing spheroidal alumina particles 6 C is the oil-drop method in which drops of an aqueous acidic alumina material gel to spheroids in falling through a water-immiscible liquid and coagulate under basic p H conditions A wide variety of oil-drop techniques have been developed in attempts to provide structural and mechanical properties that would enhance the activity and durability of alumina-supported catalysts The density, surface area, 65 W ' a fi porosity and uniformity of the spheroidal product vary greatly with the nature of the alumina feed and, along with crush strength and attrition resistance, are dependent on the conditions used in the preparation of the feed and the coagulation and gelation steps, as well as subsequent drying and calcination steps.

Internal gelation, i e gelation of the alumina by a weak base, such as 5 hexamethylenetetramine, that is added to the feed before drop formation and that releases ammonia in the heated immiscible liquid, is the most common oildrop method.

United States Patent No 3,558,508 to Keith et al describes an oil-drop method employing an external gelation technique in which gaseous ammonia is introduced 10 into the bottom of a column containing the water-immiscible liquid and coagulates the droplets by contacting their external surfaces The Keith et al process is based to a considerable extent on the use of specific alumina feed prepared by acidic hydrolysis of finely divided aluminum Spherical alumina particles may also be formed by the hydrocarbon/ammonia process described in Olechowska et al, 15 "Preparation of Spherically Shaped Alumina Oxide", International Chemical Engineering, Volume 14, No 1, pages 90-93, January, 1974 In this process, droplets of a slurry of nitric acid and dehydrated aluminum hydroxide fall through air into a column containing hydrocarbon and ammonia phases The droplets assume spheroidal shapes in passing through the water-immiscible liquid and then 20 are coagulated to firm spheroidal beads or pellets in the coagulating medium.

Similar processes utilizing pzeudosol feeds and hydrochloric acid are described in:

1 Katsobashvili et al, "Formation of Spherical Alumina and Aluminum Oxide Catalysts by the Hydrocarbon-ammonia Process 1 The Role of Electrolytes in the Formation Process", Kolloidnyi Zhurnal, Vol, 28, No 1, pp 46-50, January 25 February, 1966; 2 Katsobashvili et al, "Preparation of Mechanically Strong Alumina and Aluminum Oxide Catalysts in the Form of Spherical Granules by the HydrocarbonAmmonia Forming Method", Zhurnal Prikladnoi Khimii, Vol 39, No 11, pp.

2424-2429, November, 1966; and 30 3 Kasobashvili et al, "Formation of Spherical Alumina and Aluminum Oxide Catalysts by the Hydrocarbon-ammonia Process-Coagulational Structure Formation During the Forming Process", Kolloidnyi Zhurnal, Vol 29, No 4, pp.

503-508, July-August, 1967.

Catalysts are used to convert pollutants in automotive exhaust gases to less 35 objectionable materials Noble metals may be used as the principal catalytic components or may be present in small amounts to promote the activity of base metal systems United States Patents Nos 3,189,563 to Hauel and 3,932,309 to Graham et al show the use of noble metal catalysts for the control of automotive exhaust emissions United States Patent No 3,455,843 to Briggs et al is typical of a 40 base metal catalyst system promoted with noble metal Unpromoted base metal catalysts have been described in United States Patent No 3,322,491 by Barrett et al.

The activity and durability of an automotive exhaust catalyst is in part dependent on the location and distribution of noble metals on the support Since 45 the use of noble metal is controlled to a great extent by cost, small amounts of noble metals should be placed on the support in a manner that achieves the best overall performance over the life of the catalyst.

Several competing phenomena are involved in the surface treatment.

so Impregnating the maximum amount of the support particle provides the greatest 50 amount of impregnated surface area However, since gas velocities are high and contact times are short in an automotive exhaust system, the rate of oxidation of carbon monoxide and hydrocarbons and the reduction of nitrogen oxides arc diffusion controlled Thus, the depth of impregnation should not exceed the distance that reactants can effectively diffuse into the pore structure of the particle 55 A balance of impregnated surface area coupled with proper dispersion and accessibility should be achieved to formulate a practical catalyst.

Catalytic metal accessibility and dispersion will provide initial high catalytic activity, once the catalyst reaches operating temperature However, since significantly high amounts of hydrocarbons, carbon monoxide, and other partially 60 combusted materials are produced in exhaust gases during the initial moments of the engine start, the catalyst should have low thermal inertia in order to operate efficiency when the reaction zone is at a relatively low temperature.

A common deficiency of exhaust catalysts is decreased activity when exposed to high temperatures, mechanical vibration and poisons present in the exhaust such 65 , = ' I 1,603 462 4 1,603,462 4 as lead, phosphorus, sulfur compounds, etc, for long periods of use of up to 50,000 miles or so An effective catalyst will retain its activity through resistance to noble metal crystallite growth, poisons, crystalline phase changes, and physical degradation.

An optimum high temperature alumina catalyst support has low density and 5 high macroporosity while retaining substantial surface area and crush strength and attrition resistance Furthermore, it is stable in crystalline phases and geometric volume occupied Difficulties have been encountered in achieving the proper balance of these interrelated and sometimes competing properties and in combining an alumina support and metal impregnation techniques to provide a 10 catalytic converter capable of decreasing automotive exhaust emissions to the levels required by present and future government standards.

In our Application No 80 29047, Serial No 1,603,463, we describe and claim a process for preparing spheroidal alumina particles which comprises:

(a) commingling alumina and an acidic aqueous medium to provide a slurry; 15 (b) forming droplets of the slurry; (c) passing the droplets downwardly through air into an upper body of waterimmiscible liquid and ammonia and into a lower body comprising aqueous ammonia to form spheroidal particles; (d) aging the particles in aqueous ammonia; and 20 (e) drying and calcining the aged particles.

The alumina used in this process is preferably prepared by the process which is described and claimed in Application No Il 80/78, Serial No 1,603,461, and which comprises the following steps:

1 An aqueous solution of an aluminium salt, generally of a strong mineral 25 acid, preferably aluminium sulfate, having an A 12,0 concentration of 5 to 9 weight percent and at a temperature of 130 to 1600 F, is added to water at a temperature of to 1700 F; the amount of aluminium salt added is sufficient to adjust the p H of the mixture to 2 to 5:

2 An aqueous solution of sodium aluminate (or other alkali metal aluminate) 30 having an A 1203 concentration in excess of 16, generally 18 to 22, wzight percent and a temperature of 130 to 160 'F and a further amount of aqueous aluminium sulfate (or other aluminium salt) solution are simultaneously (but separately) added to the mixture-this precipitates alumina to form an alumina slurry Addition of alkali metal aluminate raises the p H of the mixture to a value in the approximate 35 region p H 7 to p H 8, and the p H of the slurry is maintained during the precipitation from 7 to 8 and the temperature is kept at from 140 to 180 OF and a rate of addition of the solutions is maintained during the precipitation to form intermediate boehmite-pseudoboehmite alumina; 3 The p H of the slurry is then adjusted to 9 5 to 10 5 The slurry may then, 40 optionally, be aged and the slurry is then filtered and the filter cake washed to provide a substantially pure alumina.

The present invention provides a catalyst support comprising spheroidal alumina particles and possessing a pore volume of about 0 1 to about 0 4 cubic centimeters per gram in pores of 1000 to 10,OOOA in diameter, a surface area of 45 about 80 to about 135 square meters per gram, an attrition loss of less than about %, and a compacted bulk density of about 20 to about 36 pounds per cubic foot.

This support can be prepared by the process for preparing spheroidal alumina particles described above.

It has been discovered that a catalyst comprising a catalytically active metal or 50 metal compound impregnated on the spheroidal alumina particles of this invention has excellent activity and durability in many catalytic systems It is especially suited for eliminating pollutants in automotive exhaust streams because of its quick light off and sustained activity under high temperatures and mechanical vibrations present in exhaust systems 55 The catalyst support, of this invention is in the form of spheroidal particles and is characterised by (i) a pore volume of about 0 1 to about 0 4 cubic centimeters per gram in pores of 1000 to 10,OOOA in diameter; (ii) a surface area of about 80 to about 135 square meters per gram; 60 (iii) an attrition loss of less than about 5 ', preferably less than about 2 ,'; and (iv) a compacted bulk density of about 20 to about 36 pounds per cubic foot; and may be further characterised by (v) a total pore volume of about 0 8 to about 1 7 cubic centimeters per gram; ; ' j 9 u l 1,603,462 5 (vi) a pore volume of about 0 5 to about 1 0 cubic centimeters per gram in pores of 100 to 1000 A in diameter; and (vii) a pore volume of 0 to about 0 06 cubic centimeters per gram in pores of less than 100 A in diameter; and further characterised by l 5 (viii) a volume shrinkage of less than about 6 %a, preferably less than about 4 %, 5 upon exposure to a temperature of 18000 F for 24 hours; and (ix) a crush strength of at least about 5 pounds, preferably greater than about 7 pounds.

Throughout this specification the "nitrogen pore volume" refers to the pore volume as measured by the techniques described in the article by S Brunauer, P 10 Emmett, and E Teller, J Am Chem Soc Vol 60, p 309 ( 1938) This method depends on the condensation of nitrogen into the pores, and is effective for measuring pores with pore diameters in the range of 10 to 600 A.

The surface areas referred to throughout this specification are the nitrogen

BET surface areas determined by the method also described in the Brunauer, 15 Emmett and Teller article The volume of nitrogen adsorbed is related to the surface area per unit weight of the support.

In the process to form the spheroidal alumina particles the alumina and an acidic aqueous medium, such as an aqueous solution of an acid or acid salt, are commingled to provide a slurry Preferably, an aqueous solution of a monobasic 20 mineral acid is commingled with water and the alumina to provide the slurry Use of a monobasic acid provides a homogeneous, plastic slurry with the desired viscosity Hydrochloric acid and other strong monobasic acids may be used and the support washed free of these electrolytes Aluminum nitrate may be used Nitric acid is preferred because it is decomposed and removed from the spheroids by 25 heating later in the process so that washing the spheres is not necessary In order to minimize the nitrogen oxides produced in the later states as noxious emissions, a decomposable monobasic organic acid such as acetic acid, (hereinafter represented symbolically as CH 3 COOH), formic acid, or mixtures thereof, preferably replaces a major portion of the nitric acid For example, a mixture of 30 organic acid and nitric acid in a molar ratio of about 0 5 to 5 may be employed.

Bulk density and crush strength of the spheroid product depend upon feed composition Increasing alumina and/or acid content of the feed increases these physical properties Too high a concentration of alumina and/or acid may result in spheroid fracture upon drying and too low a concentration in weak, powdery 35 spheroids Because of the gel content of the alumina powder used in preparing the feed, a minor amount of acid is sufficient to form a plastic slurry The slurry may contain about I to about 12 weight percent of a monobasic acid or mixtures thereof and the slurry generally contains about 10 to about 40, preferably about 24 to about 32 weight percent of alumina and has a molar ratio of acid to alumina of about 0 05 40 to about 0 50 The quantity of water is sufficient to yield a slurry with these acid and alumina contents Normalizing the system in relation to one mole of alumina, the inorganic acid molar ratio may vary between 0 5 to 0 03, preferably 0 06, and the organic acid molar ratio from 0 to 0 3, preferably 0 12, and the water molar ratio may be about 5 to about 50, preferably about 10 to about 20 An especially 45 preferred slurry has a molar composition of (A 1203), Q(CH 3 COOH)O 12 (HNO 3)006 (H 20)14, +, The slurry may be prepared from a single alumina composition or a blend of alumina compositions Blends are used to take advantage of some specific properties of the individual components of the blend For example, alumina filter 50 cake may be acidified with acetic acid, to about p H 6 0 prior to spray drying to reduce carbon dioxide absorption A high carbonate content in the powders may result in sphere cracking during drying Thus, 20 parts of this low carbonate alumina may be combined with 80 parts of untreated dried powder to give a blend with an acceptable carbonate level Preferably, the alumina powder and acidic 55 aqueous medium are commingled in stages by adding portions of the powder to the medium to acidify the alumina and reduce the level of CO 2 that may be present in the spray dried alumina powder For example, 80 percent of the alumina required for a given batch of product may be mixed in water which contains the desired quantities of acid After a period of mixing, the remaining 20 percent of the powder 60 is then added to the batch In addition, recycled, calcined product fines in an amount of up to about 15 percent of total alumina may be added This decreases the tendency of the product to shrink to about 2 to about 3 volume percent It also | |; ; S S ' f S r; X 1 1 1 11', '' 1 1 1 '' 1 1 -, _ 1 ' 1 makes the process more economical in that scrap product such as fines, etc, can be recycled.

Agitation and aging of the slurry provide a uniform material with a viscosity that permits proper formation of the droplets from which the spheroids with low shrinkage can be made Agitation of the slurry can be accomplished by a variety of 5 means ranging from simple hand stirring to mechanical high shear mixing Slurry aging can range from a few minutes to many days The aging time is inversely related to the energy input during mixing Thus, the alumina powder can be stirred, by hand, into the acid and water mix for 10 minutes and aged overnight to reach the proper consistency for droplet formation For example, in a specific preferred 10 method using about 10 lbs of powder, 60 percent of the powder is mixed with all of the acid and water and blended vigorously with a 1/2 H P Cowles dissolver turning a 3 inch blade at about 3500 RPM for about 2 to about 30 minutes or preferably about 15 to about 20 minutes The remaining 40 percent of the powder is then added and stirring recommenced for about 5 to about 60 minutes and preferably 15 about 30 to about 40 minutes After agitation, the slurry is aged for about I to about hours to reach the proper consistency During mixing, the p H rises and a final p H of generally about 4 0 to about 4 8, preferably about 4 3 to about 4 4, is achieved.

The viscosity of the slurry, measured immediately after the preferred blending technique, may vary between about 60 and about 300 centipoises (cps) For 2 C optimum droplet formation, slurry viscosities of about 200 to about 1600 cps, preferably about 800 to about 1200 cps are desirable Viscosities as high as 2000 cps may be used but the slurries are difficult to pump.

Under actual operating conditions in a plant, there might be occasions in which a slurry may have to wait for long periods of time prior to further processing 2 ' Under these conditions, the viscosity of the system may climb above the pumpable range Such a thickened slurry need not be wasted It still can be used by following any of the following two procedures:

The thick slurry may be diluted with controlled amounts of water and strongly agitated for short periods of time This will result in a sharp decrease of the 3 C viscosity and will bring the system into the pumpable range.

The thick slurry may be mixed with a freshly prepared slurry which will exhibit a low viscosity between about 60 and about 300 cps The resulting mixture will have a viscosity in the pumpable range and can be used in the process.

Both of these remedial steps can be practiced without adversely affecting the 35 properties of the finished product nor the subsequent processing steps.

The viscosity of slurries referred to in these specifications, examples and claims, is the viscosity as measured with a Brookfield viscometer.

The spheroidal particles are formed by gelation in an organic phase and an aqueous phase Droplets of the aged slurry are formed in air above a column which 4 C contains an upper body of water-immiscible liquid and ammonia and a lower body of an aqueous alkaline coagulating agent The drops assume spheroidal shapes in passing through the upper phase and then are coagulated into firm spheroidal particles in the lower phase The ammonia in the upper phase gels the droplet exterior layers sufficiently to allow the spheroidal shape to be retained as the 4 ' droplets cross the liquid interface and enter the lower phase Excessive interfacial tension between the phases may result in retention of the droplets in the organic phase and possibly their deformation In such cases, a small quantity of surfactant, for example, about 0 05 to about 0 5, preferably 0 1 to about 0 2 volume percent of the upper body, is placed at the interface and permits the spheres to penetrate it 5 ( easily Liquinox ), a detergent sold by Alconox, Inc, New York, N Y, and other such surfactants may be employed.

The water-immiscible liquid will have a specific gravity lower than water, preferably lower than about 0 95, and can be, for example, any of the mineral oils or their mixtures The organic liquid should not permit the droplets to fall too 55 rapidly which may inhibit proper sphere formation Furthermore, it should not exhibit high interface surface tension which may hold up and deform the particles.

Examples of suitable mineral oils, include kerosene, toluene, heavy naphtha, light gas oil, paraffin oil, and other lubricating oils, coal tar oils, and the like Kerosene is preferred because it is inexpensive, commercially available, non-toxic and has a 6 C relatively high flash point.

The organic liquid should be capable of dissolving small amounts of anhydrous gaseous ammonia or be capable of forming suspensions containing trace amounts of water which contain dissolved ammonia An essential requirement of the process is that the organic phase contain sufficient, but small, amounts of a base, preferably 6 ' 1,603,462 7 1,603,462 7 ammonia, in order to be able to effect the partial neutralization and gelation of the outer layers of the falling droplets The rate of introduction of ammoniainto the organic liquid should be sufficient to reach an operating concentration in which firm particles will be formed in the short time span of fall However, the ammonia concentration should not be so high as to cause essentially instantaneous gelation 5 of the slurry droplets as they enter the organic liquid Under these conditions, the droplets will gel into misformed particles since they have not had sufficient time of fall to allow their surface tension to spheroidize the droplet Furthermore, high concentrations of ammonia in the upper regions of the organic liquid will cause evaporation of gaseous ammonia into the air pocket where the nozzles are located 10 Excessive ammonia concentration in this region may cause premature gelation of the droplets prior to the point of separation from the nozzle This is very undesirable because premature gelation in the nozzle will cause plugging and malfunction of the delivery system Ammonia is a preferred coagulation agent because it produces good spheroids, exhibits a convenient solubility, and may be 15 conveniently introduced into the lower portion of the organic liquid In a preferred embodiment, the organic liquid is contacted with anhydrous gaseous ammonia in a separate apparatus called the ammoniator, and circulated through the column In such an event, the organic liquid from the ammoniator is introduced in the lower portion of the organic phase in the column and it flows upwardly through the 20 column establishing a counter current flow with the falling droplets The organic liquid is removed at the top of the column and returned to the ammoniator for replenishing with added ammonia.

Under steady state conditions, an ammonia concentration gradient develops within the organic phase of the column The gradient is caused by the reaction of 25 the falling acidic alumina slurry droplets with the ascending ammonia carried by the organic phase Because of the lower ammonia concentration in the upper portions of the column, the droplets have time to shape into spheroids before they gradually gel as they descend The ammonia concentration in the organic liquid may be determined by titration with hydrochloric acid to a bromthymol blue 30 endpoint and may be maintained between about 0 01 to about 1 0, preferably about 0.04 to about 0 07, weight percent Lower concentrations generally result in flattened spheroids, and higher concentrations in deformations such as tail formation.

The length of the column can vary widely and will usually be from about 3 to 35 feet in height The organic phase may generally comprise about 1/3 to about 2/3 of the column length and the coagulation phase the remainder.

The aqueous medium may contain any substance capable of inducing gelation and having an appropriate specific gravity, i e lower than the specific gravity of the slurry droplets This permits the spheres to pass through it Alkaline aqueous 40 solutions such as sodium hydroxide, sodium carbonate, or ammonia can be used as the coagulating medium The preferred medium is an aqueous solution of ammonia, because it and its neutralization products are easily removed from the spheroids in later processing steps Washing is not necessary to remove the ammonium residue as it would be to remove a sodium residue The ammonia 45 concentration in the aqueous phase may be about 0 5 to 28 4 weight percent preferably about 1 0 to about 4 0 weight percent During prolonged use, ammonium nitrate and acetate may be formed and build up to steady state levels in the aqueous phase These are products of the neutralization reaction occurring during sphere gelation Their steady state concentration will be dependent upon 50 the concentrations of the acids in the alumina slurry feed In the development of this invention ammonium acetate and ammonium nitrate were added to the aqueous ammonia phase to simulate the effects of eventual steady state values of these salts.

For the preferred slurry composition, the concentrations used were typically about 1 3 and about 0 8 weight percent respectively 55 Under continuous operation, ammonia must also be constantly added to the aqueous phase to replace that used in gelation of the spheres In a preferred embodiment of this invention, the aqueous phase is circulated between the column and an ammoniator tank This tank also serves as a reservoir with a batch collection system to take up aqueous ammonia solution displaced from the column as spheres 60 fill up the collection vessel The aqueous phase is removed from the column to maintain a constant interface level In a continuous sphere take-off system, the reservoir feature of the aqueous phase ammoniator would not be needed Either type of collection system can be used.

The cross sectional area of the column is dependent upon the number of 65 8 1,603,462 8 droplet nozzles used For one nozzle, a one inch diameter column provides approximately 5 cm 2 of cross-sectional area, which is sufficient to keep the uncoagulated droplets from hitting the column walls and smearing and sticking on the walls A four inch diameter column provides enough cross-sectional area for up to about 16 to 20 nozzles to permit the droplets to fall independently through the 5 column without contacting each other or the walls.

In one embodiment of a suitable column, the aged slurry is pumped into a pressurized multiple orifice feed distributor that is located at the top of the oil column and contains a multiplicity of nozzles positioned about 1/2 inch above the organic liquid The pressure of the feed distributor is dependent upon the slurry 10 viscosity Pressures of about 0 1 to about 15 p s i g are normally used The feed distributor pressure regulates the droplet formation rate The latter varies from about 10 to about 250 droplets per minute with a preferred rate being about 140 to about 180 drops per minute A distributor pressure of about 1 5 to about 2 5 p s i g.

gives the desired droplet rate when the slurry viscosity is in the range of about 800 15 to about 1200 cps The nozzles employed can vary in diameter to give spheroidal particles of the desired size For example, a 0 11 inch internal diameter nozzle will produce spheroids of a diameter of about 1/8 inch Preferably, an air flow is provided around the nozzles to keep ammonia vapor from prematurely gelling the droplets The droplets of slurry are formed in air at the nozzle tips and fall through 20 air into the body of water-immiscible liquid When the drops of slurry initially contact the immiscible liquid, they are usually lens-shaped As the drops fall through ammonia-treated organic liquid, they gradually become spheroidal particles which are set into this shape by the coagulating ammonia and harden further in the lower aqueous ammonia phase 25 The particles are then aged in aqueous ammonia with a concentration of about 0.5 to 28 4 weight percent, preferably the same concentration as in the column The particles develop additional hardness so that they are not deformed during subsequent transfer and processing steps In general, the particles may be aged from about 30 minutes to about 48 hours, preferably about I to 3 hours 30 The particles are then drained and dried Forced draft drying to about 2100 to about 4000 F for about 2 to 4 hours may be advantageously employed although other drying methods may also be used In a preferred drying method, the drying is done in a period of under 3 hours by programming the temperature to climb gradually and uniformly to about 3000 F The amount of air used may normally vary 35 between about 400 and 600 standard cubic feet per pound of A 1203 contained in the wet spheroids Under certain circumstances, some of the air may be recirculated in order to control the humidity of the drying medium The spheres are usually spread over a retaining perforated surface or screen at thicknesses ranging from I to 6 inches preferably 2 to 4 inches A slight shrinkage usually occurs during drying but 40 the spheroids retain their shape and integrity.

Deviations from the prescribed conditions of preparation of starting raw materials may often result in significant changes in the products obtained.

Excessive powder particle size, crystallinity, or level of impurities may result in cracking and fracturing during drying On the other hand excessive levels of gel in 45 the powder, or pseudoboehmite may result in excessive shrinkage and densification upon drying which can also lead to cracking Alumina compositions other than the product of our invention which are suitable for spheroid formation will generally have a boehmite or pseudoboehmite crystalline structure, preferably microcrystalline, a nitrogen pore volume of 0 4 to 0 6 cm 3/g, surface areas in 50 excess of 50 m 2/g and will contain amorphous gel.

The dried spheroid product is then treated at high temperatures to convert the crystalline alumina hydrate and amorphous gel components to a transition alumina.

This may be done by batch or continuous calcination by contacting the product with hot gases which may be either indirectly heated gases or the combustion 55 products of ordinary fuels with air Regardless of the particular method used, the product is calcined at specific temperature ranges depending on the particular transition alumina desired.

For example, to obtain a gamma type alumina, the product may be conveniently calcined at temperatures of about 10000 F to about 1500 'F For 60 applications which require high temperature stability while retaining high surface area and porosity, the target material may be theta alumina A predominantly theta alumina product may be obtained by calcination at about 17500 to about 19500 F.

preferably about 1800 to about 19000 F for periods of from about 30 minutes to t l Do -9 1,603,462 9 about 3 hours, preferably from about I hour to about 2 hours For automotive exhaust catalysts, the high temperature treatment step is often called stabilization.

The catalyst support that comprises the spheroidal alumina particles and that "-: is obtained after stabilization generally has the following range of properties:

,-:: -a 5Approximate 5 General Property Range Surface Area (m 2/g) 80-135 Conmpacted Bulk Density (Ibs /ft 3) 20-36 Tota Pore Volume (cm 3/g) 0 8-1 7 10 Pore Size Distribution (cm 3/g) Below 100 OA 0-0 06 1000 A O 5-1 0 1000 10,000 A 0 1-0 4 Above I 0,OOOA 0-0 4 15 Crush Strength (Ibs -force) 5-15 Volume Shrinkage ( 0) 6 Attrition Loss () 0-5 Mesh Size -4 + 10 However, when the preferred starting raw materials are used under the 20 preferred conditions of preparation, the property ranges become:

Typical Property Range Surface Area (m 2/g) 90-120 Compacted Bulk Density (Ibs /ft 3) 26-32 25 Total Pore Volume (cm 3/g) 0 9-1 2 Pore Size Distribution (cm 3/g) Below 100 A 0-0 04 oo1000 A O 6-0 9 1000 10 O,OOOA O 2-0 3 30 Above 10,000 A 0-0 3 Crush Strength (Ibs -force) 7-12 Volume Shrinkage (%) 2-4 Attrition Loss (%) 0-2 Mesh Size -5 + 7 35 The surface areas are nitrogen BET surface areas and the other above specified properties were determined by the following methods These methods may be also applied to the finished catalysts.

Compacted Bulk Density A given weight of activated spheroids is placed in a graduated cylinder 40 sufficient to contain same within its graduated volume "Activated" as used herein means treated at 320 O F in a forced draft oven for 16 hours prior to the testing This activation insures that all materials are tested under the same conditions The cylinder is then vibrated until all settling ceases and a constant volume is obtained.

The weight of sample occupying a unit volume is then calculated 45 Total Specific Pore Volume A given weight of activated spheroids is placed in a small container (for example, a vial) Using a micropipette filled with water, the said sample is titrated with water until all of the pores are filled and the endpoint of titration occurs at incipient wetness of the surface These measurements are consistent with total 50 porosities calculated from the equation:

f I P= _ D p in which:

P=total specific porosity (cm 3/g) f=volume packing fraction (for spheroids typically ( 0 64 + 0 04) 55 D=compacted bulk density (g /cm 3) p=crystal density of skeleton alumina (g /cm 3) (typically between 3 0 and 3 3 g./cm 3 for transition aluminas) :

1,603,462 10 Mercury Pore Size Distribution The pore size distribution within the activated spheroidal particle is determined by mercury porosimetry The mercury intrusion technique is based on the principle that the smaller a given pore the greater will be the mercury pressure t ' ' 5 required to force mercury into that pore Thus, if an evacuated sample is exposed 5 to the mercury and pressure is applied incrementally with the reading of the mercury volume disappearance at each increment, the pore size distribution can be determined The relationship between the pressure and the smallest pore through which mercury will pass at the pressure is given by the equation:

-2 acos O r 10 P 1 where r=the pore radius a=surface tension 0 =contact angle P=pressure 15 Using pressures up to 60,000 p s i g and a contact angle of 1400, the range of pore diameters encompassed is 35-l O,OOOA.

Average Crush Strength Crush strength is determined by placing the spheroidal particle between two parallel plates of a testing machine such as the Pfizer Hardness Tester Model 20 TMI 41-33, manufactured by Charles Pfizer and Co, Inc, 630 Flushing Avenue, Brooklyn, New York The plates are slowly brought together by hand pressure The amount of force required to crush the particle is registered on a dial which has been calibrated in pounds force A sufficient number (for example, 50) of particles is crushed in order to get a statistically significant estimate for the total population 25 The average is calculated from the individual results.

Shrinkage A given amount of particles is placed in a graduated cylinder and vibrated until no further settling occurs, as is done in determining Compacted Bulk Density This sample is then placed in a muffle furnace at 180001 F for 24 hours At the end of this 30 exposure, its volume is again measured after vibration until no further settling occurs The loss in volume after heating is calculated, based on the original volume, and reported as percent shrinkage.

Attrition Loss A set volume ( 60 cc) of material to be tested is placed in an inverted 35 Erlenmeyer flask of special construction which is connected to a metal orifice inlet.

A large (one inch) outlet covered with 14-mesh screening is located on the flat side (bottom) of the flask High velocity dry nitrogen gas is passed through the inlet orifice causing the particles to: (I) circulate over one another thus causing attrition, and ( 2) impact themselves in the top section of the flask thus breaking down as a 40 function of strength The material is tested for five minutes and the remaining particles are weighed The loss in weight after testing expressed as percent of the initial charge is designated the attrition loss.

The nitrogen flow will be in the range of about 3 5 and 4 0 cubic feet per minute, depending upon the density of the material The flow rate must be 45 sufficient for the particles to strike the top section of the flask The fines produced by attrition are carried out of the flask by the nitrogen flow thus causing a loss in weight of the original material charged.

The alumina and sphere formation conditions of the present invention provide spheroidal alumina particles with a highly unexpected and uniquely desirable 50 combination of properties The spheroids have a total pore volume ranging from about 0 8 to about 1 7 cm 3/g While this is a high total pore volume, in itself it is not exceptional What makes this pore volume exceptional is the size distribution of the pores which make up this volume and high temperature stability of this volume A large fraction of the volume is made up of macropores (> 1000 A) Most of the rest 55 of the pores are in the 100-1000 A range There are very few micropores (<IOOA).

This type of distribution is important for catalytic activity and stability In a t A; ,Iv> a,;z;; ; heterogeneous process, catalytic activity is highly dependent upon the rate of diffusion of reactants to the catalyst sites and of reaction products away from the sites Thus, reaction processess in a catalyst containing a large amount of macroporosity are less diffusion dependent However, the macropores account for only a small fraction of the sample surface area The intermediate size pores provide 5 the surface area required for catalytic activity This surface area has two components, namely, that required by the catalytically active clusters themselves and that required to keep the clusters separated If the clusters are allowed to fuse together, their catalytic surface area and consequently the catalyst activity will decrease Microporosity of course, provides a very large surface area, but, this 10 does not necessarily provide good catalytic activity Diffusion of reactants and/or products may be the rate controlling factor Micropores can be closed over by sintering occurring during catalyst operation or by deposition of poisons such as lead compounds in an auto exhaust system In either case, the activity of the catalyst in the closed micropores would be lost 15 The surface area of the product spheroids is high, but is not unusually high.

Surface areas range from about 350 m 2/g to about 500 m 2/g for spheroids heated to 1000 F and drop to about 80 m 2/g to about 135 m '/g for thermally stabilized spheroids at 1800-19000 F What is important, however, is that most of the surface area is associated with intermediate size pores and not with micropores 20 This preferred porosity distribution and its pore volume stability are a direct result of the unique combination of properties in the alumina powder used to make the spheroids In particular, its purity and high ratio of crystalline material to amorphous gel aid in minimizing microporosity.

These properties also help to account for the high temperature stability of the 25 spheroids The spheroids exhibit low volume shrinkage, preferably less than about 4 % They retain the transition alumina structure Alpha alumina is not detected even at temperatures of 19500 F It is well known that impurities act as sintering aids Thus, high impurity levels can promote shrinkage and alpha alumina formation A high gel content also leads to alpha alumina formation at high 30 temperatures A high microporosity can result in high volume shrinkage as micropores are closed off during sintering.

The spheroids also have an uncommon combination of low bulk density and relatively high crush strength Low bulk density is essential for quick light off, i e.

high initial catalytic activity The crystallinity of the alumina compositions is a 35 contributing factor to both the low bulk density and high crush stren-gth.

The low attrition loss exhibited by the spheroids is a direct consequence of their shape and strong structure The smooth surface will not attrit as readily as irregular surfaces which exhibit corners and/or edges Also, the gelation process produces a coherent uniform particle rather than a layered particle which results 40 from some mechanical balling processes A mechanically formed particle may delaminate during an attrition process.

Another feature of this invention is the close control of the spheroid size In a given batch, greater than 95 % of the spheres will be within one mesh size, such as -5 + 6 or -5 + 7 Measurement with a micrometer shows that the spheroids are even 45 more closely sized There is only about a 0 015 inch variation in the major or minor axis of the spheroids Thus, a controlled distribution of sphere sizes can be obtained by using the proper distribution of nozzle sizes This can aid in controlling the pressure differential across a packed catalyst bed which is an important factor in auto emissions catalyst devices 50 The properties of the spheroid product, when taken in toto define a unique particle which makes a superior catalyst support.

The support of this invention is characterized by low density, high degree of macroporosity, high crush strength, high temperature shrink resistance, good attrition resistance and controlled size and shape 55 Preparation of Catalyst Having the Support A catalyst comprising the support of this invention impregnated with a catalytically effective amount of at least one catalytically active metal or metal compound is highly effective in many catalytic systems particularly those that operate at high temperatures Although the alumina itself may be active as a 60 That is, 95 of the spheres pass a sieve with an aperture of 4 mm but are retained on a sieve with an aperture of 3 55 mm.

-;; , ,; 1 1 I,603,462 1 1 catalyst, it is usually impregnated with a suitable catalytic material and activated to promote its activity The selection of the catalytic material, its amount, and the impregnation and activation procedures will depend on the nature of the reaction the catalyst is employed in Preferably, the catalytically active metal is platinum group metal selected from the group consisting of platinum, palladium, ruthenium, 5 irridium, rhodium, osmium, and combinations thereof.

In order to possess the high initial and sustained activity necessary to meet the increasingly stringent emission controls required by state and federal laws in the United States of America, and by the national laws of other countries, the catalytic agents in a catalyst of this invention are distributed in particular positions within 10 the catalyst particle The catalytic agents which are typically used in automobile exhaust catalysts are the platinum group metals Because of the great cost of these metals, it is uneconomical to use large quantities Hence, it is important to position the metals in the most efficient as well as the most strategic manner Suitable platinum group metals include, for example, platinum, palladium, rhodium, 15 ruthenium, iridium, and osmium, as well as combinations thereof.

The use of platinum group metals in automobile exhaust catalysis has been constrained by the natural abundance of these metals Since a large fraction of the world supply occurs in South Africa with the major metals being platinum, palladium and rhodium which occur naturally in the approximate proportions of 68 20 parts platinum, 27 parts palladium and 5 parts rhodium, the majority of work has been with the platinum group metals in these proportions Typically, the total noble metal content of automotive exhaust catalysts, based on the weight of the catalyst, is about 0 005 to about 1 00 weight percent but preferably is from about 0 03 to 0 30 weight percent to be both economically as well as technically feasible Of course, 25 the optimum combination of metals is that which imparts specific performance benefits to the catalyst such as a higher level of palladium for more rapid oxidation of carbon monoxide and better thermal stability, or more platinum for greater poison resistance and better durability of hydrocarbon oxidation, or increased rhodium concentration for improved conversion of nitrogen oxides to nitrogen Catalysts 30 prepared according to this invention may contain only one of the platinum group metals, but usually contain several The metal ratios may vary over a wide range of values but preferably the catalyst contains platinum and palladium in a weight ratio of about 5 parts to about 2 parts, respectively, or platinum, palladium and rhodium in a proportion to about 68 parts, 27 parts and 5 parts, respectively 35 A variety of noble metal compounds are well-known and have been documented in the literature The type of compound to be used depends largely on the nature of the surface of the support and the resulting interaction with it For example, anionic or cationic forms of platinum may be introduced into the support by using H Pt CI 6 or Pt(NH 3)4 (NO 3)2, respectively 40 The noble metal compound to be used should be introduced into the support such that once it is decomposed into the active form (metal or metal oxide) it will be highly dispersed and positioned in a specific location in the catalyst particle We have found that there are preferred locations for the metals as well as preferred distribution of one metal with respect to another for a high degree of initial activity 45 and which is especially important for sustained durability The ability to control the location and dispersion depends on the noble metal compound used We have found that of particular utility are those complexes described in U S Patent 3,932,309 to Graham et al In this method, the catalyst is prepared by impregnating the support with sulfite-treated platinum and palladium salt solutions These sulfito 50 complexes when applied to the support decompose to provide a high degree of dispersion By varying the cationic form of the complex, for example, NH 4, or H form, the depth of impregnation of the metal within the particle can be varied.

We have found that for high initial, and more importantly sustained activity, the noble metals should be positioned such that about 50 of the total noble metal 55 surface area is deeper than about 50 microns With regard to the specific location of the noble metals, we have found that the preferred location of the noble metals should be such that about 50 4 of the active metals are located deeper than about microns Determination of noble metal surface area is carried out by the hydrogen titration of oxygen The detailed method for carrying out this 60 determination is described by D E Mears and R C Hansford, J of Caratlvis Vol.

9, 125-134 ( 1967) Elemental analysis is carried out using conventional analytical procedures In order to determine the surface areas and noble metal concentration at particular depths within the catalyst particle, a method called the chloroform attrition method is employed This method involves the agitating of a certain 65 I 1,603,462 1 2 weight of catalyst in a liquid (chloroform) for a specified length of time dependent upon the amount of surface to be attrited off The attrited material is separated from the unattrited remainder, dried and weighed Knowing the initial dimensions of the catalyst particles as well as their geometry and weight, the depth removed can be determined This attrited material is analyzed for its noble metal surface area and noble metal content From these data the noble metal surface area and noble metal contents can be calculated as a function of depth into the catalyst particle.

The activity and durability of the catalysts were ultimately tested by several methods Bench scale activity testing is carried out by a method which simulates the cold start that a catalyst experiences on a vehicle Once the catalyst heats up, after a period of time it will reach essentially a steady state condition whereupon the efficiencies attain a level dependent upon the intrinsic activity of the catalyst.

In the bench scale test, a 13 cm 3 sample is contacted with sufficient total gas flow rate to achieve a gas hourly space velocity of 38,000 hr -' The simulated exhaust contains 1700 ppm carbon as propane, 4 5 volume percent carbon monoxide with the balance made up by nitrogen The preheated gas, if containing no oxidizable species, would heat up the bed to 7000 F However, due to the presence of carbon monoxide and hydrocarbon, the temperature climb in the bed is accelerated due to the exothermic oxidation reactions In this test the parameters of importance are the rapidity of lightoff as measured by the time to reach 50 O conversions of carbon monoxide and hydrocarbon and the carbon monoxide and hydrocarbon conversion efficiencies which occur when the catalyst reaches essentially a steady state condition The catalysts of this invention exhibit very short times to reach 50 percent conversions and have very high conversions of hydrocarbon and carbon monoxide.

In order to assess the durability of catalysts upon exposure to fuel additives which contain lead, sulfur, and phosphorus along with the phenomena of varying temperature conditions, the catalysts are aged on a pulse flame combustor system.

In this system a fuel doped with all the poison precursors one finds in current fuelsis burned resulting in the deposition of poisons such as lead, sulfur and phosphorus compounds on the catalyst The fuel which is hexane doped with tetramethyl lead antiknock mix, trimethyl phosphite and thiophene contains the equivalent of O 23 g lead per gallon, 0 02 g phosphorus per gallon and 0 03 weight percent sulfur The fuel flow is 15 ml per hour Nitrogen and oxygen are mixed in amounts of 2000 and 500 standard cubic centimeters per minute, respectively An additional 50 standard cubic centimeters of oxygen is added after the point of combustion to ensure an oxidizing environment A detailed explanation of the operation of a pulse flame combustor operation was presented by K Otto, R A Dalla Betta and H C Yao, J.

Air Pollution Control Association, Vol 24, No 6, pp 596-600, June 1974 A 13 cm 3 sample is exposed to repetitive temperature cycling consisting of one hour at 13000 F and 2 hours at 1000 F Temperatures all at the average bed temperatures.

During prolonged aging periods of up to 500 hours the sample is periodically checked for activity by removing it from the pulsator unit and testing it on the bench scale activity unit The small scale activity and aging tests are valuable in screening many catalysts, however, the ultimate test is the full size vehicle or engine dynamometer evaluation A detailed description of an engine dynamometer test is reported by D M Herod, M V Nelson and W M Wang, Societv of Automotive Engineers, Paper No 730557, May 1973 This test is similar to the bench scale activity test reported earlier The ambient temperature catalyst contained in a full size converter is contacted with the hot exhaust from a closely controlled engine The conversions of carbon monoxide and hydrocarbon and nitrogen oxides are monitored as a function of time Correlations have been developed which allowed prediction of how the catalyst would perform if tested in an actual CVS test run by the Federal Test Procedure as detailed in the Federal Register of July, 1970 and as modified by the instructions in the Federal Register of July 2, 1971.

Durability testing of full size converter charges of catalysts is carried out by the method outlined by J P Cassassa and D G Beyerlin, Society of Automotive Engineers, Paper No 730558, May, 1973.

Once the catalyst has passed laboratory and engine dynamometer activity and durability testing, it then undergoes fleet testing on standard production type vehicles The vehicles are tested according to the Federal Test Procedure noted previously and hereby incorporated by reference The procedure is designed to determine the hydrocarbon, carbon monoxide and oxides of nitrogen in gas emissions from an automobile while simulating the average trip in an urban area of I 1,603,462 14 1,603,462 14 7-1/2 miles from a cold start The test consists of engine start up and vehicle operation on a chassis dynamometer through a specified driving schedule consisting of a total of 1,371 seconds A proportionate part of the diluted gas emissions is collected continuously for a subsequent analysis using a constant volume sampler 5 The dynamometer run consists of two tests, a cold start test after a minimum of 12 hours wait, and a hot start test with a 10-minute wait between the two tests.

Engine start up and operation over a driving schedule and engine shutdown constitute the complete cold start test Engine start up and operation over the first 505 seconds of the driving schedule complete the hot start test 10 The engine emissions are diluted with air to a constant volume and a portion sampled in each test Composite samples are collected in bags and analyzed for hydrocarbons, carbon monoxide, carbon dioxide and oxides of nitrogen Parallel samples of diluted air are similarly analyzed for hydrocarbons, 15 carbon monoxides and oxides of nitrogen Vehicle aging of the catalyst is carried I out following the driving schedule as set forth in Appendix "D" schedule in the Federal Test Procedures, noted previously This schedule consists of eleven 3 7 mile laps of stop and go driving with lap speeds varying 30-70 mph, repeated for 50,000 miles The average speed is 29 mph Periodically the vehicles are tested on a chassis dynamometer to assess the durability of the emission control system 2 ( Catalysts designed for three-way control of carbon monoxide, hydrocarbons and nitrogen oxides require a support of good physical integrity because of the variation in reducing or oxidizing nature of the exhaust environment Furthermore, they especially require a support having low density and a high degree of macroporosity because of the inherent difficulty in achieving good carbon 2 ' monoxide removal and the quick lightoff needed for good all-round performance.

Since rhodium is typically the catalytic agent added for improved threeway control and coupled with its being a rather scarce resource, it must be used efficiently and placed strategically Similar to the case of oxidation catalysts, threeway catalysts have high initial activity as well as good sustained performance when the active metals are properly positioned We have found that for combined high initial activity and sustained catalytic durability, the noble metals should be located such that about 50,' of the total noble metal lies deeper than about 75 microns.

Laboratory determination of three-way catalyst activity is made by contacting 13 cm 3 of the catalyst with sufficient gas flow to reach a gas hourly space velocity 3 of 40,000 hr-1 The gas composition consists of 1 % carbon monoxide, 250 ppm hydrocarbon consisting of a 3/1 mixture of propylene and propane, 0 34 o' hydrogen, 1000 ppm NO, 12 % carbon dioxide, 13 % water and varied oxygen contents to change the exhaust gas environment from reducing to oxidizing in nature Nitrogen is added as the balance The measure ( 0) of reducing or oxidizing 4 nature of the exhaust gas environment is given by the following:

Actual Concentration of Oxygen in the Feed Composition Oxygen Concentration Required for Stoichiometry The catalyst is evaluated at various values of O and at various steady state temperatures.

Bench scale durability is carried out in essentially the same way as in the case 4 of oxidation catalysts Periodic activity checks are run during the aging schedule.

Full scale engine or chassis dynamometer evaluation in technically and economically feasible systems is carried out with close control of air/fuel ratios.

Aging is carried out in systems with similar control of the exhaust gas environment to which the catalyst is exposed Periodic checks of three-way performance are 5 made during the aging schedule.

The catalysts of this invention have been tested by a number of methods which have been developed by the automobile manufacturers to measure catalytic performance One particular test measures conversion efficiency at a temperature of 1000 F and a gab hourly space velocity of 75,000 hour The test is particularly 5 discriminative in determining the relative ability of catalysts to oxidize hydrocarbons Below are typical ranges of performance for highly preferred catalysts of our invention compared to those achieved by typical catalysts of current manufacture.

1,603,462 15 HC Efficiency Aged 24 Hours Fresh 18000 F.

Catalysts of this Invention 64-650 % 40-44 ' 5 Catalysts of Current Manufacture (Typical) 38-410, about 35 % HC Efficiency=hydrocarbon conversion efficiency at steady state Another useful test that has been used to differentiate catalysts of our invention from those of current manufacture is one that measures the temperature at which fifty percent of the carbon monoxide in the test gas mixture is oxidized at a 10 gas hourly space velocity of 1400 hour-' Low temperature results mean high activity Below are typical ranges of performance of highly preferred catalysts of our invention compared to those of typical catalysts of current manufacture:

Temperature for 50 ' Carbon Monoxide Conversion 15 Aged 24 hrs.

Fresh at 1800 F.

Catalyts of our Invention 225-230 o F 265-3000 F.

Catalyts of Current Manufacture 300-315 F 350-3700 F.

Because of the low density and macroporosity of the spheroids of this 20 invention, the catalytic agents, for example, the platinum group metals, when applied to specific locations and in specific distributions within the catalyst can be utilized very efficiently Because of this efficient usage, the catalyst need not be loaded with noble metals to a level which exceeds its economic limits Hence, the ranges of total noble metal loading are: 25 Broad Range Normally Total Noble Metal Loading Weight 0 005-1 0 0 03-0 30 Percent The particular choice of catalytic agents used depends upon the performance characteristics desired in the system Principally, the noble metals used in 30 automobile exhaust emission control are platinum, palladium, and rhodium and mixtures thereof.

The approximate ranges of these principal components are:

Normally Broad Range Preferred 35 Platinum ( O of total) 0-100 % 65-75 %W Palladium ( 8/ of total) 0-100 %o 25-35 % Rhodium (/ of total) 0-20 ' 5-15 % Because of the high degree of macroporosity built into the support, the noble metals may be positioned deeper than in typical catalysts of current manufacture, 40 and as a result they are highly dispersed and more resistant to crystallite growth.

Hence, the catalysts are characterized as having high and stable metal surface areas The following ranges distinguish advanced catalysts of our invention from typical catalysts of current manufacture:

Noble Metal Surface Area (micromoles of H 2 per gram 45 of catalyst with a metal loading of 0 332 troy ounces/ft 3) Aged 24 Hours Fresh 18000 F 50 Catalysts of this invention Broad Typical 38-7 6 0 5-0 7 Catalysts of Current Manufacture Broad Typical 0 6-3 5 0 0-0 1 r zpm; t;e orC; 1,603,462 The catalysts of our invention are further characterized by the specific depths to which the catalytic agents are deposited There are pronounced differences between the catalysts of our invention compared to typical catalysts of current manufacture as noted below:

Approximate Maximum Depth of Noble Metal Penetration Catalysts of our Invention Broad Typical Preferred Catalysts of Current Manufacture Broad Typical 150-400 microns 150-250 microns 30-125 microns Because of the specific performance characteristics that we build into our catalyst by changing the distribution and location of the various noble metals, those distributions and locations are distinguishing characteristics of the catalyst of our invention Although the overall proportions of the catalyst as a whole may be fixed, the distribution of these components is specifically located in the catalysts of our invention The following indicates the preferred ranges of depths and distributions that characterize the catalysts of our invention:

Approximate Maximum Depth of Penetration Platinum Broad Typical Preferred Palladium Broad Typical Preferred Rhodium Broad Typical Preferred 125-400 microns 125-250 microns 125-400 microns 125-250 microns 125-250 microns 125-200 microns Catalysts of our invention are characterized by the following performance characteristics which distinguish them from typical catalysts of current manufacture.

Fresh Laboratory Dynamic Heat-up Activity CO Typical Parameter HC Efficiency Typical Catalysts of our Invention 40-50 sec 75-95 O Catalysts of Current Manufacture > 60 55-75,, t 50 CO=time in seconds for 50 ' carbon monoxide conversion.

Laboratory Dynamic Bench Heat-up Activity after 500 hours of Pulsator Aging CO HC Efficiency Parameter Typical Typical Catalysts of our Invention Catalysts of Current Manufacture 65-95 sec.

> 135 sec.

35-50 <, l 0-206 <, 0 120 O The following examples illustrate specific embodiments of the invention.

Examples I to 7 are included to illustrate the preparation of preferred starting materials.

EXAMPLE I

This example illustrates the preparation of an alumina composition.

Alumina trihydrate was completely dissolved in sodium hydroxide to provide a I ? ri 17 1,603,462 17 sodium aluminate solution containing 20 percent A 1203 and having a Na 2 O/AI 203 mole ratio of 1 40 495 grams of water were added to a reaction vessel and then 631 milliliters of 50 percent sodium hydroxide solution were added This volume of - sodium hydroxide solution corresponded to 966 grams at the specific gravity of the solution of 1 53 g /cm 3 The mixture was stirred gently and heated to 200 F A total 5 of 672 grams of alumina trihydrate was added gradually over a period of 30 minutes During the addition of the alumina trihydrate, the mixture was heated to a gentle boiling and stirred slowly Gentle boiling and stirring were then continued for another 60 minutes or until all the trihydrate was dissolved.

Heating was stopped and the mixture cooled with stirring to 140 F 10 The specific gravity and temperature of the sodium aluminate solution were adjusted to 1 428 g /cm 3 and 130 O F respectively by adding 290 grams of water at a temperature of 140 F and stirring the mixture 2016 grams of the solution were used for the preparation of the alumina.

2286 grams of an aluminum sulfate solution containing 7 percent A 1203 and 15 having a specific gravity of 1 27 g /cm 3 at 25 C and a SO 4 =/A 1203 mole ratio of 3.01 were prepared by dissolving 1373 grams of aluminum sulfate crystals in 1963 grams of water.

The sodium aluminate solution and the aluminum sulfate solution were heated to 145 F A heel of 3160 grams of water was placed in a strike tank, the agitator was 20 started, and the heel heated to 155 F.

The heel was acidified to a p H of 3 5 by the addition of 6 milliliters of aluminum sulfate at an addition rate of 36 ml /minute and aged for 5 minutes At the conclusion of the aging period, the flow of sodium aluminate was started at a rate of 28 ml /minute Within 5 seconds, the flow of aluminum sulfate was resumed 25 at 36 ml /minute and maintained constant through the 50 minute strike phase The flow of sodium aluminate was adjusted as needed to maintain the p H of the reaction mixture at 7 4 The strike temperature was maintained at 163 F by heating the strike tank.

In 50 minutes, all of the aluminum sulfate solution had been added and 317 30 grams of sodium aluminate remained.

At the conclusion of the strike, the p H of the reaction mixture was increased to 10 0 by adding 29 more grams of sodium aluminate solution The final molar tatio of Na 2 O to SO 4 was 1 00 The solution was stabilized by aging for 30 minutes at a constant temperature of 163 F 35 After aging, the reaction mixture was filtered and washed For every gram of alumina in the mixture, 50 grams of wash water were used A standard filtrationwash test was defined as follows Reaction slurry ( 600 ml) was filtered in an 8 inch diameter crock using Retel filter cloth, material no 80, at 10 inches of vacuum It was washed with 2 5 liters of water The filtration time was 2 1 minutes and the 40 filter cake was 7 mm thick.

The filter cake was reslurried at 15 % solids and spray-dried at an outlet temperature of 250 F to a powder having a total volatiles (T V) content of 27 5 %, as measured by loss on ignition at 1850 F The dried powder was calcined at 1850 F for I hour 45 The properties of the dry product and the calcined product are shown in Table 1.

TABLE I

Dry Powder Wt % Na O 0 02 50 Wt % 563 0 20 Wt % T V 27 5 Agglomerate Size 21 5 u Bulk Density 24 1 lbs /ft 3 X-Ray Phases 55 (boehmite-pseudoboehmite intermediate) no alpha or beta trihydrate phases present peak for l 020 l crystallographic plane falls at d spacing of 6 37 A 60 Calcined Powder at 1850 F For I Hou, N 2 Surface Area 130 m 2/g.

N 2 Pore Volume, > 600 A 0 72 cm 3/g.

total 0 95 cm 3/g.

'::'-;,,',: ' x,,;:':7,'':,:; 7 ';:::;,;',::,:::',,:; - -:

X-Ray Phases Pore Size Distribution 1,603,462 TABLE I (Cont) theta alumina, no alpha alumina present A nitrogen PSD measurement showed that all the pores were greater than 100 A diameter and that 50 % of the pores were in the 100-200 A diameter range.

EXAMPLE 2

Given below in Table 2 is a summary of results of 13 runs conditions described in Example 1.

using the process TABLE 2

Properties Average No of Runs Wt A 12 O Run (lbs) Strike Ratio-Na 2 O/SO= 4 Standard Filtration Test (min) 13 I 0.93 2.4 Spray Dried Powder Wt /% Na O Wt % so= Wt % T V.

Bulk Density (Ibs /ft 3) N 2 Surface Area at 750 F for 30 minutes (m 2/g) N 2 Pore Volume at 750 F for 30 minutes (cm 3/g) < 600 A X-Ray Calcined Powder at 1850 F for 1 Hour Surface Area (m 2/g) Pore Volume (cm '/g) Total < 600 A X-Ray 0.03 0.19 27.9 24.0 420 0.82 Intermediate BoehmitePseudoboehmite 131 1.01 0.73 Theta alumina, no alpha alumina present EXAMPLE 3

Given below in Table 3 is a summary showing the results for the blended products of six large scale runs The process was the same as in Example I except that 195 lbs of alumina (dry basis) were made per run Equipment size and amounts of material were scaled up proportionately The results were the same as in laboratory scale runs showing that the process could be readily scaled up.

TABLE 3

Spray Dried Powder Wt % Na 2 O Wt % SOWt % CO, Wt % T V.

Bulk Density (lbs /ft 3) N 2 Surface Area at 750 F (m 2/g) N 2 Pore Volume at 750 F (cm 3/g) < 600 A X-Ray Phases 0.059 0.31 1.37 29.6 30.0 413.

0.77 Intermediate BoehmitePseudoboehmite i{ t St I,>Nw OU t -N nkn;T X.

'', ' ' 7 _ 1 ' 1 ' V, ' '1'Il '1 ,-p V11;, ', i 1 -, ' ' ' -, ' ',, #"d' ', 'c', 1 ' 1 7 1,603,462 TABLE III (Cont) 5:-', ":: ' 5 a,,1; 5 Calcined Powder-1850 F /l Hr.

N 2 Surface Area (m 2/g) N 2 Pore Volume (cm 3/g) Total < 600 A X-Ray Phases 131 0.97 0.70 Theta alumina, no alpha alumina present EXAMPLE 4

The process conditions shown in Example I were important to obtain an easily filterable, pure product In runs where the process conditions of Example I were employed except that the reaction temperature and time were varied, the following results were obtained.

TABLE 4

Reaction Temperature Reaction Time Standard Filtration Time Test (min) Wt % SO; 750 F 1200 F 1630 F.

min 25 min 50 min.

(Example I)

2.8 9 0 2 1 9.5 O 19 0 20 Thus, a decrease in process temperature led to an increase in sulfate content.

A decrease in process time leads to an increase in filtration time.

EXAMPLE 5

This example illustrates the treatment of a washed alumina filter cake prepared in accordance with the procedure of Example 3 with acetic acid before spray drying The acetic acid has the effect of decreasing the absorption of carbon dioxide during spray drying In each run glacial acetic acid was added to the filter cake to a p H of 6 0 and the mixture agitated The spray dried product contained 3 8 percent acetic acid and 64 5 percent alumina This represents 0 I moles of acetate ion for each mole of alumina.

The properties of the alumina products of Example 3 and this Example are shown in Table 7 The carbon dioxide content is 0 76 % compared to 1 37 % present in the alumina of Example 3.

TABLE 5

Spray Dried Powder Example 3

Example 5

Wt % Na 2 O Wt % SO= Wt % Solids (A 1203) Wt % CO 2 X-Ray Calcined Powder 1850 F /f Hr.

N 2 Surface Area (m 2/g) N 2 Pore Volume (cm 3/g) < 600 A X-Ray 0.059 0.31 63.8 1.37 Intermediate BoehmitePseudoboehmite 131 0.70 theta alumina, no alpha alumina present 0.053 0.50 64.5 0.76 Intermediate Boehmite Pseudoboehmite 118 theta alumina, no alpha alumina present EXAMPLE 6

In order to illustrate the relative proportions of crystalline material and amorphous material in the alumina, samples of the alumina of Example I and aluminas A and B that exhibit lower and higher degrees of crystallinity respectively were slurried with deionized water at A 1203 concentrations of 100 g A 1203 (dry 1 1 l 1 :; ' ' i basis) in one liter of water Potentiometric titrations of each slurry were slowly conducted at a rate of addition of 1 1 N sulfuric acid of I mi /minute over the p H range of 8 3 to 4 0 in which alumina is insoluble.

= ' TABLE.

TABLE 6

Volume of l IN H 2 SO 4 Solution Required 5 to Reach Indicated p H Example 1 A B d l 0201 spacing 6 37 A 6 56 A 6 1 IA midpoint width 1 78 A 1 98 A 0 18 A of peak l 020 l 10 p H 8.3 0 0 0 7.0 36 47 14 6.0 90 120 25 5 0 117 156 36 15 4.0 150 193 47 The results show that the preferred alumina compositions required intermediate amounts of acid to effect the same p H change and thus had a gel content intermediate between A and B. EXAMPLE 7 20 -.: The degree of crystallinity of the alumina compositions was further demonstrated by X-ray diffraction measurements of the development of beta ;, alumina trihydrate on alkaline aging and heating 100 gram samples (dry basis) of alumina A as shown in Table 6 and the alumina prepared in Example 3 as shown in Table 3 were slurried in 250 milliliters of deionized water and brought to p H 10 by 25 the addition of IN Na OH solution The time and temperatures of aging and the height of the high and low intensity X-ray peaks of alumina A for beta trihydrate are shown in Table 7 No detectable beta trihydrate was present in the alumina composition of Example 3 under the same conditions of aging and heating as alumina A 30 TABLE 7

4.72 A Peak 4 35 A Peak Time/Temperature Height (mm) Height (mm) 18 hrs /50 C 8 8 24 hrs /50 C 11 8 35 41 hrs /50 C 10 12 4 hrs /90 C 12 14 21 hrs /90 C 18 14 The ease of formation of beta trihydrate under alkaline conditions of Sample A indicated a higher gel content than in the alumina of Example 3 40 EXAMPLE 8 :. In a series of runs, five cubic feet of spheroidal alumina particles with an ; average bulk density of 28 pounds per cubic foot were prepared using a mixture of ( 1) a blend of the products of the runs of Example 3 and ( 2) the acetic acid treated alumina of Example 5 An 80/20 mix of plain alumina to acetate alumina was used 45 and slurried in nitric acid, acetic acid, and water The composition of the mixture was:

Example 3 Alumina ( 63 8 Wt % A 1203) 1919 g.

Example 5 Alumina ( 64 5 Wt % A 1203 3 8 Wt % H 3 CCOOH) 475 g 50 1.5 M HNO 3 600 ml.

1.5 M CH 3 COOH 1000 ml.

Water 1780 ml.

Nominal Composition of above mix.

(A 1203), oo(CH 3 COOH) 012 (HNO 3)o o(H 2 O),s 34 55 ':-,,,: -':

-,,:, -: 3 ',:l,-:' -' '- "-4 " xl:-': ', ; f ,' r: >" 1,603,462 21 1,603,462 21 The liquids were mixed together in a five gallon bucket and blended with the alumina of Example 3 using a Cowles Dissolver with a three inch diameter blade turning at 3500 R P M A 20 minute blending was used The acetate alumina was then added and the slurry was blended for another 20 minutes Viscosity of the slurry immediately after blending wdl&-78 cps 7 as measured with a Brookfield 5 viscometer The initial viscosity varied between 60 to 100 cps The slurry was aged to a viscosity of 500 to 1000 cps before being used for sphere forming After aging, the p H's of the slurry varied between 41 and 4 5 in the runs.

After aging, the alumina slurry was pumped to a pressurized feed tank that was 5 inches in diameter and 4 inches high The slurry was continuously circulated 10 between the feed tank and a reservoir tank to maintain the viscosity of the slurry.

The alumina slurry feed flowed under air pressure of O 5 to 1 5 p s i g from the feed tank to the nozzle holder The droplet formation rate varied between 140 and 170 drops/minute The nozzle holder could hold up to nineteen 2 7 mm internal diameter nozzles in a regular array 7 to 14 nozzles were used per run and the extra 15 openings in the nozzle holder were used as spares in case any of the original nozzles clogged The nozzle holder contained air channels to provide a linear air flow of cm /minute around the nozzle tips and prevent ammonia vapor from prematurely gelling the alumina droplets.

The 2 7 mm internal diameter of the droplet nozzle was selected to give about 20 1/8 inch diameter (minor axis) calcined spheres The lip thickness was 0 6 mm The 3.3 mm nozzle holes opened to 1/2 inch diameter, 1/2 inch long cylindrical holes cut in the bottom of the holder The ends of the stainless steel nozzles were recessed 1/8 inch from the bottom of the holder and the bottom of the nozzle holder was 1/4 inch above the organic phase of the column 25 With a slurry viscosity of 700 cps and a feed pressure of 1/2 p s i g a droplet rate of 170 drops per nozzle per minute could be maintained using seven nozzles It took 1-1/4 hours to form the slurry batch into droplets.

A glass sphere-forming column was employed The column was 9 feet high anid 4 inches in diameter The column was filled with kerosene (no 1 grade) and 28 % 30 aqueous ammonia The top six feet of the column contained the kerosene The remainder of the column was filled with the aqueous ammonia The aqueous ammonia was mixed with 1 3 wt % ammonium acetate and 0 83 wt % ammonium nitrate as measured under steady state operating conditions The kerosene was ammoniated to a concentration range of 0 03-0 08 wt % ammonia The kerosene 35 also contained 0 2 volume % Liquinox.

A glass column 4 feet high by 3 inches diameter was used to ammoniate the kerosene The column was half filled with ceramic saddles Kerosene was pumped from the top of the sphere forming column at a rate of approximately one liter per minute to the top of the ammoniating column Ammonia gas flowed into the 40 bottom of this column The ceramic saddles broke up the stream of ammonia bubbles permitting a more efficient ammoniation of the kerosene Ammoniated kerosene was pumped from the bottom of the ammoniator column to the bottom of the kerosene layer in the spheroid forming column An ammonia concentration gradient existed within the kerosene phase of the spheroid forming column The 45 top of the kerosene phase had the least ammonia The ammonia concentration at the top of the kerosene phase was maintained between 0 03 and 0 08 wt % The concentration was determined by titration with HCI to a bromthymol blue endpoint.

A batch collection system was used An 8 liter bottle was connected to the 50 bottom of the spheroid forming column by detachable clamps A one inch diameter ball valve was used to seal off the bottom of the column when the collection bottle was detached The collection bottle was filled with 28 % aqueous ammonia An overflow reservoir was connected to the collection bottle to catch the aqueous ammonia displaced by the spheroids When the collection bottle was full the 55 spheroids were poured into a plastic basin where they were aged in contact with 28 % aqueous ammonia for one hour prior to drying.

A forced air drying oven was used The spheroids were dried in nesting baskets with a 20 mesh stainless steel screen bottom The top of each basket was open to the bottom of the basket above it The top basket was covered The bottom basket 60 contained a charge of previously formed spheroids saturated with water Because the sides of the baskets were solid the flow of water vapor during drying was down and out through the bottom of the stack of baskets A humid drying atmosphere was maintained in this manner to prevent spheroid cracking Drying temperature '.

1 1 1 '17 1 1 1', 1 1 Z __': ' =' --, 22 1,603,462 22 was 260 F A 30 ft long, gas fired tunnel kiln with a 14 inch square opening was used to calcine the spheroids at 1900 F for one hour.

A summary of the run conditions and properties of the calcined spheroids for :';:: this example are given in Table 8.

? 5 A summary of the properties of the approximately three cubic foot blend of 5 calcined spheroids formed in a series of runs by the conditions of this example are shown in Table 9 The spheroid bulk density and average crush strength were relatively uniform Attrition and shrinkage were low.

TABLE 8

Slurry 10 Wt % Al 203; Nominal 26 5 Actual 27 6 Blend Time (Min) 20 + 20 p H 4 39 Aging Before Run (Hours) 4 5 15 Run Viscosity; Initial (cps) 700 1900 F Calcination Bulk Density (Ibs /ft 3) 28 2 Crush Strength (lbs) 9 0 Range-High 11 5 20 -Low 7 5 Major Axis (mils) 148 Minor Axis (mils) 130 Major/Minor Axis ratio 1 14 Water was evaporated during the mixing process 25 TABLE 9

Weight (lbs) 78 7 Volume (ft 3) 2 74Bulk Density (lbs /ft 3) 28 7 Average Crush Strength (lbs) 10 5 30 % Attrition 0 5 % Shrinkage 3 5 Sphericity (Major Axis/Minor Axis) 1 13 Average Diameter (mils) 135 N 2 Surface Area (m 2/g) 107 35 X-Ray Theta alumina, no alpha aluminaEXAMPLE 9

In Table 10, the slurry and spheroid forming properties of threeconventional 40 alumina powders are compared with those of three different powders produced by the powder formation process of Example 3 The process described in Example 8 was used to produce the slurries and spheroids.

With both C and D alpha alumina monohydrate powders, it was necessary to use a lower solids content in the slurry With a higher solids content than those 45 used, the slurries set solid in a few minutes Also, with the C alumina, it was necessary to use a higher alumina-acid ratio At the standard ratio, the slurry set up while blending.

The C and D aluminas resulted in high bulk density spheroids Although the same size nozzles were used in all cases, these two aluminas formed spheroids 50 which were much smaller than the spheroids formed by the alumina compositions of this invention.

A crystalline alpha alumina monohydrate E was made by heating Alcoa C-30 D alpha alumina trihydrate to 300 F for 4 hours A slurry made at the standard alumina-acid ratio had a p H of 3 2 The solids settled out immediately after 55 blending When the alumina-acid ratio was doubled (to 1/0 09) the p H was 3 9, but the solids still settled out immediately after blending.

Y, Powder Powder Properties %/ Total Volatiles Crystallographic Form as is Crystallographic Form, 1900 F.

for I Hour N 2 Surface Area, 1900 F for 1 Hour (m 2/g) N 2 Pore Volume, < 600 A, (cm 3/g) at 1900 F for I hour Slurry Forming % Solids Mole Ratio; A 12 J 0/Acids Slurry p H Initial Viscosity, (cps) Time to reach 1000 cps.

Sphere Properties After 1900 F Calcination Crystallographic Form N 2 Surface Area (m 2/g) N 2 Pore Volume, < 600 A (cm 3/g) Bulk Density Average Crush Strength (lbs) Major Axis Diameter (mils) Minor Axis Diameter (mils) TABLE 10

Example 3 Powders 25.6 24.2 C 24.4 25 1 Alpha Alumina Monohydrate D 21.5 Theta Alumina 0.70 27.6 1/0 18 4.4 84 4 hours 123 0.59 28.2 9.0 148 122 0.67 27.5 1/0 18 4.5 320 I hour 0.60 28.7 10.0 141 118 133 0.63 32.0 1/0 18 4.1 400 2 hours Theta Alum 104 0.61 28.1 13.4 121 0.40 19.1 1/0 09 4.5 400 min.

ina 103 0.39 55.6 18.9 102 124 0.54 23.4 1/0 18 4.5 220 1 hour E 21.3 Kappa Alumina 0.1 1/0 18 3.2 Settled out 108 0.53 52.4 21.3 116 t) -0 % C (V 0 % k) I'J '-1 A.' V 1 I' -'1 4 ' t 11 '1 1,603,462 The following examples illustrate the preparation and testing of the catalyst of this ifnvention.

EXAMPLE 10

An alumina slurry feed was made from an alumina powder which had the following characteristics:

0.08 wt % Na 2 O 0.43 wt % SO0.095 wt /o Ca O 0.022 wt % Mg O 29 4 wt % Total Volatiles 0.85 cm 3/g N Pore Volume 300 m V/g N 2 Surface Area at 1000 F.

X-ray diffraction shows alpha alumina monohydrate with the l 0201 reflection occurring at 6 6 A The slurry had the following composition:

17.5 wt % alumina 4.2 wt % nitric acid ( 0 38 moles HNOJ/mole A 1203) The slurry was formed by hand stirring It was aged for 2 days Spheroids were formed in a 1-inch diameter column Kerosene was the water immiscible phase.

The aqueous phase contained about 28 weight percent ammonia Three 175 g.

batches were made and combined The samples were calcined at 1000 F for 3 hours The properties of the spheroids were:

Bulk Density:

Crush Strength:

Water Pore Volume:

Size: 28.6 pcf.

13.2 lbs.

1.08 cm 3/g.

*-6 + 7 mesh After calcination at 1850 F for I hour, it had:

Bulk Density: 34 3 pcf.

Crush Strength: 10 0 lbs.

Noble metal catalysts were prepared on the concentration was 0 04 troy oz /260 cu in in a I to dynamic heat-up oxidation activity test, the following calcined substrates The 3 Pt/Pd weight ratio In a results were obtained:

Fresh 24 Hrs /1800 F.

CO Index HC Efficiency, % 0.677 94.8 0.925 78.9 These tests were conducted in accordance with the procedure of U S Patent No 3,850,847 to Graham et al, except that the simulated exhaust gas contained 1700 ppm carbon as propane.

The sample had excellent catalytic activity and stability for both carbon -40 monoxide and hydrocarbon conversion 40 EXAMPLE 11

Spheroidal alumina particles prepared in accordance with the procedure of Example 8 had the properties shown in Table 11.

1900 grams of these particles were impregnated to incipient wetness with a solution prepared as follows: 45 S Ol was bubbled into 800 ml deionized water for 17 minutes at I m.

mole/minute after which 4 213 ml of Pd (NO 3)2 solution containing 105 mg.

palladium per ml was added The resulting solution is yellowish green indicating complexing of the palladium to a degree of 4 moles SO 2/g atom palladium.

A solution of ammonium platinum sulfito salt, (NH 4)6 Pt(SO 3)4 x H 2 O, was 50 prepared by dissolving 3 678 g having a platinum content of 30 67 % in 700 cc.

water.

The palladium solution was then added to the platinum solution The total volume was then increased to 1938 ml by the addition of additional deionized water The solution was applied via a stream to the rotating support Once 55 impregnation was complete, the support was placed on screens and oven dried at 320 F (forced draft) After overnight drying it was activated at 800 F for I hour in air.

't, j ItI , 1 ',, 1 ';,t ' ', ' ' 'J,;,:-, 1,'111; '1 'Z1-11 1 ','I 1 1 1,603,462 TABLE 1

Bulk Density Crush Strength (lbs) Attrition % Sphericity (Major Axis/Minor Axis) Surface Area (m 2/g) X-Ray 28.0 9.7 0.20 1.1 104 0 theta alumina, no alpha alumina present The bench and engine dynamometer, and vehicle test results for this catalyst are shown in Table 12 and the bench activity during aging on the pulse flame combustor are shown in Table 13 The catalyst exhibited high hydrocarbon conversion efficiency in the High Space Velocity Test It also was determined to have very low temperatures for 50 % carbon monoxide and hydrocarbon conversion in the static bench test The dynamometer aging data indicated good performance after 1000 hours of engine aging The catalyst was also aged on a vehicle in fleet tests and the results were quite similar to those obtained in aging on a engine dynamometer As observed in full scale engine tests, the light off (t 50 co) parameter as determined after extensive pulsator aging was quite good.

In the tables which follow, the abbreviations used have the following meaning:

HSV=High Space Velocity GHSV=Gas hourly space velocity Cat =Catalyst of the present invention Std =Standard or reference catalyst typical of current commercial catalyst in use in the U S A.

CVS=Constant volume sampling as per standard Fed test procedure.

ND=Not determined HC=Hydrocarbons TABLE 12

Bench Test Results Fresh Aged HSV Static HSV Static HC CO H C CO -C CO HC CO Cat 65 100 230 225 40 100 275 265 Std 38 100 311 313 35 100 369 370 Conversion efficiency at 1000 F 75,000 GHSV-' % conversion temperature, 1400 GHSV-' Aged 24 hours at 1800 F Dynamometer Data-Fresh Time to 50 % Cofinv.

Seconds 600 Sec Eff Pred CVS Eff.

HC CO HC CO HC CO Cat 37 29 98 100 92 83 Std 57 44 93 99 87 81 Dynamometer Data-Aged 1000 Hours Time to 50 % Cony.

Seconds 600 Sec Eff Pred CVS Eff.

HC CO HC CO HC CO Cat.

Std.

97 72 79 97 78 87 76 95 74 77 r ,, l' -'i' l Z eÀ 2 Z' -25 : 5 Five Car Fleet Test-Details Vehicles:

Mileage Accummulation:

cars-350 V-8, 4 BBL with M-Air Rotate converters every 5000 miles, (Every converter will be on each car twice) 5 Programmed chassis dynamometer 1 f 977 Schedule Five Car Test Fleet 10,000 Mile Catalyst Evaluations Predicted CVS Data HC Conversion Eff, Veh Aged Dyno, Aged 87 CO Conversion Eff, Veh Aged Dyno Aged 78 81 TABLE 13

Bench Activity During Aging on Pulse Flame Combustor Total Aging Hours 0 69.5 136 5 203 5 Activity with Propane Feed t 50 CO (secs) 54.3 66.0 57.5 76.8 t'50 HC (secs) 95.4 ND ND HC Eff.

(%) 79.9 57.2 51.1 38.8 Activity with Propylene Feed HC (secs) 54.0 69.6 76.8 96.9 Fuel: 0 23 g Pb/gal; 0 02 g P/gal; 0 03 % S Catalyst underwent 152 cycles of 2 hours at 1000 F and I hour at 1300 F.

Temperature is average axial bed temperature EXAMPLE 12

Eleven 1300 gram batches of spheroidal alumina particles prepared in accordance with the procedure of Example 8 and having the properties shown in Table 14 were impregnated as follows first with palladium, second with platinum.

TABLE 14

Average Bulk Density (lbs /ft 3) Crush Strength (lbs) Average High Low Attrition (%) Sphericity (Major Axis/Minor Axis) Average Diameter (mils) Surface Area (m 2/g) X-Ray 28.8 8.7 12.2 17.0 9.0 1.1 1.14 113 theta alumina, no alpha alumina The palladium solution was prepared by dissolving SO, at 2 m moles/min for minutes in 800 ml deionized water, after which 4 63 ml Pd(NO 3)2 at 105 mg /ml.

palladium solution was added To this 2 00 grams of dibasic ammonium citrate were added, then the solution volume raised to 1277 ml It was impregnated to incipient wetness, then dried on screens at 320 F for a minimum of I 1/2 hours, then dried at 500 F overnight.

The platinum was then applied from a solution prepared by dissolving 4 077 g.

%? >:n '"" À,<, -I X, ;, r? w Cat.

Std.

4 C C 1,603,462 ',,, ^'_;, 1:

A 1 1 27 1603,462 27 (NH 4) Pt( 503)4 x H 20 at 30 67 % Pt in 800 ml of water and then raising the impregnation volume to 1277 milliliters The impregnated support was dried on screens at 320 F, then activated at 800 F for 1 hour in air.

The bench activity during aging on the pulse flame combustor is shown in Table 15.

TABLE 15

Bench Activity During Aging on Pulse Flame Combustor Activity with Propane Feed t 50 CO t 50 HC HC Eff CO Eff.

(secs) (secs) (%) ( /) 45.3 61.2 63.6 66.7 92.4 251 4 377 4 81.1 59.5 55.2 46.4 99.5 99.4 99.3 98.2 Activity with Propylene Feed t'50 HC (secs) 42.3 62.7 82.5 90.3 Fuel: 0 23 g Pb/gal; 0 02 g P/gal; 0 03 % S Catalyst underwent 170 cycles of 2 hours at Temperature is average axial bed temperature.

16.

1000 F and 1 hour at 1300 F.

The bench and dynamometer test results for this catalyst are shown in Table TABLE 16

Bench Test Results Fresh Aged HSV Static HSV Static HC CO HC CO HC CO HC CO Cat 64 100 230 225 44 100 305 300 Std 41 100 302 303 35 100 351 351 Conversion efficiency at 1000 F 75,000 GHSV-' %/ conversion temperature, 1400 GHSV-1 Aged 24 hours at 1800 F Dynamometer Data-Fresh Time to 50 % Conv.

Seconds 600 Sec Eff Pred CVS Eff.

HC CO HC CO HC CO Cat 22 17 94 100 91 83 Std 37 30 93 99 89 83 Dynamometer Data-Aged 1000 Hours Time to 50 % Conv.

Seconds 600 Sec Eff Pred CVS Eff.

HC CO HC CO HC CO Cat.

Std.

59 79 97 78 73 91 79 78 All of the test results (bench and engine) show the catalyst to be excellent in fresh performance and very good in its ability to retain its activity.

Total Aging Hours 0 69.5 0 210 5 1, Y ' V .

f ' li, ' 1 - ',:"'-4 1, ll''1" 1 1 1 1 # 1.603462 11 11 ' "' 1 r ''I' ' ' 1.603 462 EXAMPLE 13

The noble metal penetration in the catalyst of Example 12 was determined by the chloroform attrition method and the results are shown in Table 17.

These results indicate very deep penetration of both the platinum and the palladium The platinum was higher at the surface to ensure good hydrocarbon performance, whereas the palladium was distributed very uniformly to ensure good light off retention.

TABLE 17

Cumulative Noble Metal S.A % 29 44 54 64 68 73 77 Cumulative Depth Attrited (microns) 18 81 101 146 173 Cumulative Platinum (%) 14 28 39 51 59 72 Cumulative Palladium (%) 4 II 18 32 37 EXAMPLE 14 À A catalyst was prepared by impregnating spheroidal alumina particles that were prepared in accordance with the procedure of Example 8 and that had the properties shown in Table 18.

TABLE 18

Bulk Density (lbs /ft) Crush Strength (lbs) Sphericity (Major Axis/Minor Axis) Surface Area (m 2/g) X-Ray 27.1 11.4 1.32 112 theta alumina, no alpha alumina present cc ( 43 7 g) of the particles were impregnated to incipient wetness with a solution prepared by dissolving 59 mg of (NH 4)6 Pd(SO 3)4, x H 2 O (containing 17.84 % palladium) and 90 mg of (NH 4)B Pt(SO 3), x H 2 O (containing 29 28 %o platinum) in 42 ml water After impregnation, the catalyst was dried on a screen at 320 F in a forced draft oven It was then activated at 800 F in air for one hour.

This catalyst exhibited outstanding fresh and pulsator aged performance In particular, it had good light off, for example, as shown by very stable '50 CO values.

TABLE 19

Bench Activity During Aging on Pulse Flame Combustor Activity with Propane Feed CO (secs) 46.2 58.5 62.1 57.6 t 50 HC (secs) 66.0 8 7 337 5 HC Eff.

(%) 89.5 69.0 64.7 56.1 Activity with Propylene Feed t 50 HC (secs) 45.9 57.9 74.4 63.6 Fuel: 0 23 g Pb/gal; 0 02 g P/gal; O 03 % S Catalyst underwent 160 cycles of 2 hours at 1000 F and 1 hour at 1300 F.

Temperature is average axial bed temperature.

; i ,,,-, t, , i -Ae a,<ev Total Aging Hours 0 137 5 206 5 1 -_ 11 1 1 1 1 r: ;: , 1,603,462 EXAMPLE 15

A catalyst was prepared by impregnating spheroidal alumina particles that were prepared in accordance with the procedure of Example 8 and that had the properties shown in Table 20.

TABLE 20

Bulk Density (lbs /ft 3) Crush Strength (lbs) Sphericity (Major Axis/Minor Axis) Attrition loss (%) Surface Area (m 2/g) X-Ray 29.7 9.2 1.16 0.25 % theta alumina, no alpha alumina present cc ( 49 02 grams) of the particles were impregnated to incipient wetness with a solution prepared by bubbling SO 2 at I m mole/min for 20 seconds into 10 milliliters of water, to which was added 0 100 ml of Pd(NO 3)2 at 105 mg Pd per ml.

To this solution was then added 0 682 ml of acid platinum sulfito complex prepared by cation exchanging of (NH 4)6 Pt(SO 3)4 x H 2 O which contained 38 6 mg.

Pt/ml Total impregnation volume was increased to 42 ml After impregnation the sample was placed on a screen and forced draft oven dried at 320 F The sample was activated at 800 F in air for I hour.

The catalyst contains 0 05 oz /ft 3 total noble metals at 5/2 Pt/Pd ratio The penetration depth was 25 to 50 microns.

The bench activity data for this catalyst are shown in Table 21 The catalyst performance is good but not nearly as is observed on catalysts with deeper penetrations.

TABLE 21

Bench Activity During Aging on Pulse Flame Combustor Activity with Propane Feed t 50 CO (secs) 52.0 90.6 99.5 Total Aging Hours 0 146 t 50 HC (secs) 72.9 HC Eff.

(%/) 90.6 30.9 25.2 CO Eff.

(%) 99.5 99.3 98.9 Activity with Propylene Feed t 50 HC (secs) 52.5 121 2 116 1 HC Eff.

99.5 97.2 97.2 CO Eff.

(%) 99.7 99.2 99.2 Test terminated due to rapid loss in propane efficiency Fuel: 0 23 g Pb/gal; 0 02 g P/gal; 0 03 % S Catalyst underwent 49 cycles of 2 hours at 1000 F and 1 hour at 1300 F.

Temperature is average axial bed temperature EXAMPLE 16

A three way catalyst was prepared on the support described in Table 22.

TABLE 22

Average Bulk Density (lbs /ft 3 Crush Strength (lbs) Attrition % Sphericity (Major Axis/Minor Axis) X-Ray 29.8 9.3 0.14 1.23 theta alumina, no alpha alumina present , ''' -, Ml, ME,,,,,, ' r '-vi; A,; -: 9 : Total Aging Hours 0 146 ' ' ' ' 7 ' p _ ' ? 1: ',,, -1 1 1: 11 1 1 -, i "1,'', kf I e 1,'1 1,, 1 1 1,603,462 30 Two batches of 1300 g (= 2 724 liters) of support were impregnated as follows:

The support was sprayed to 1/2 of incipient wetness using an atomizing nozzle with a solution prepared by bubbling SO 2 into 400 milliliters of water for 6 12 minutes at 2 m moles SO 2/min To this 2 757 milliliters of Pd (NO 3)2 solution at 105 -5 mg Pd/ml were added Then 1 284 g ammonium citrate (dibasic) were added and 5 volume increased to 610 milliliters total Immediately after palladium application, a solution prepared by dissolving 2 2005 g (NH 4)6 Pt( 503)4 x H 20 @ 32 B 8 % platinum in 400 milliliters of water and then diluting to 610 milliliters was sprayed to the remainder of incipient wetness It was dried at 320 F for 2 hours and then at 500 F for I hour It was then sprayed to 95 % of incipient wetness with a solution 10 prepared by diluting acid rhodium sulfite solution which was prepared by cation exchanging (NH)6 Rh(SO 3)4 x H 20 using a cation exchange resin 4 24 milliliters of acid rhodium sulfite @ 50 65 mg rhodium per ml were diluted to 1160 milliliters and then sprayed on the support The impregnated catalyst was dried at 320 F and was then activated at 600 F for I hour 15 Resultant total noble metal loading was 0 332 oz total noble metal per cubic foot of catalyst.

In Table 23, the results of three way catalyst testing are described Considering the very small amount of rhodium present, the conversion of nitrogen oxides to nitrogen was quite high which is attributed to the proper positioning of the rhodium 2 C in the spheroidal particles.

TABLE 23

N Ox Conversion Efficiencies at Approx 40,000 GHSV 0-0 70 O = 0 95 Bed Temperature 750 O F 900 F 1050 F 750 F 900 F 1050 F 2:

Fresh NOX-N, 64 8 75 0 87 5 97 7 97 7 97 7 NOX (total) 98 9 98 9 100 98 9 98 9 98 9 Feed: 1 % CO, 250 ppm HC(C 3 H 6/CH 8 = 3/l), 0 34 % H 2, 1,000 ppm NO, 12 o CO 2, 13 % H 2 O, varying concentrations of 02, balance N 2 3 ( is a measure of air/fuel ratio defined as Actual concentration of 0, in the feed 4 = 0, concentration required for stoichiometry EXAMPLE 17

A catalyst was prepared on the spheroidal alumina particles of Table 18 by impregnating platinum and palladium on different particles 35 Two components were prepared; platinum on alumina, and palladium on alumina They were prepared to provide a blend of 42 31 by weight Pd component and 57 69 % by weight Pt component that gave an equal atom loading on each support The total noble metal loading is 0 332 oz per cubic foot.

The palladium particles were prepared by incipient wetness impregnation of 4 ( cc ( 43 7 g) of the support with a solution of (NH 4)6 Pd(SO 3)4 x H 20 17 84 ",, palladium dissolved in sufficient water to give a total volume of 42 milliliters The impregnated support was dried at 320 F and activated for I hour at 800 F.

The platinum particles were prepared by incipient wetness impregnation of 100 cc ( 43 7 g) of the support with a solution of (NH 4)6 Pt(SO 3)4 x H 20 2 29 28 ', 4:

platinum dissolved in sufficient water to give a total volume of 42 milliliters The impregnated support was dried at 320 F and activated for I hour at 800 F.

The fresh activity and activity after 24 hours at 1800 F is shown in Table 24 and the pulsator aging in Table 25 The fresh and thermal aged activities were excellent Pulsator aged performance was quite good 5 ( TABLE 24

Fresh CO (seconds) 44 HC efficiency (),,) 90 24 hours 5 ' at 1800 F.

CO (seconds) 61 HC Efficiency (o) 70 g,-, ': -;,-?; 7:' ' '; ';" '"' 'v: - ':;:'o' 'r ' " '=:;::, ,'?; ? ' 31 1,603,462 31 TABLE 25

Bench Activity During Aging on Pulse Flame Combustor Activity with Propane Feed Activity with Propylene Feed Total t 50 CO t 50 HC HC Eff Co Eff t 50 HC HC Eff Co Eff.

Hours (secs) (secs) (%) (%) (secs) ( 0) ( 0/o) 5 0 44 4 63 0 90 3 99 2 38 1 99 2 99 2 69 56 2 138 0 72 6 99 3 58 5 98 9 99 2 138 5 60 0 184 8 61 4 99 3 57 0 98 7 99 4 209 0 63 7 531 3 52 6 99 3 69 9 98 0 99 3 Fuel: 0 23 gPb/gal: 0 02 gP/gal; 0 03 % S 10 Catalyst Underwent 70 cycles of 2 hours at 1000 F and I hour at 1300 F.

Temperature is average axial bed temperature EXAMPLE 18

The spheroids discussed in Example 8 were measured for nitrogen pore size and surface area distributions The technique used is described by E V Ballou, 15 and O K Doolen in their article, Automatic Apparatus for Determination of Nitrogen Adsorption and Desorption Isotherms, published in Analytical Chemistry, Volume 32, pp 532-536 (April, 1960) The equipment used for this determination was an Aminco Adsorptomat manufactured by American Instrument Company of Silver Spring, Maryland 20 The nitrogen BET surface area of this material was 120 m 2/g with the following distribution:

Approximate of Cumulative Nitrogen Pore Diameter Surface Area to 25 (A) Indicated Diameter 600 1 3 % O 500 1 6 % 400 2 3 % 300 5 0 % 30 16 3 % 46 40 about 1000 It is obvious from these data that the vast majority of the pores were in the intermediate range of 100-1000 A More specifically, over 80 % of the pores were 35 between 100 and 200 A, and that no surface area was detected by this technique below pores of 100 A in diameter.

Claims (19)

1. WHAT WE CLAIM IS:-

   I A catalyst support comprising spheroidal alumina particles and possessing a pore volume of about 0 I to about 0 4 cubic centimeters per gram in pores of 1000 40 to 10,OOOA in diameter, a surface area of about 80 to about 135 square meters per gram, an attrition loss of less than about 5/, and a compacted bulk density of about to about 36 pounds per cubic foot.
2. 2 A support according to claim I which possesses a total pore volume of about 0 8 to about 1 7 cubic centimeters per gram, a pore volume of about 0 5 to about 1 0 45 cubic centimeters per gram in pores of 100 to 1000 A in diameter, and a pore volume of 0 to about 0 06 cubic centimeters per gram in pores of less than 100 OA in diameter.
3. 3 A support according to claim I or 2 which possesses a volume shrinkage of less than about 6 o upon exposure to a temperature of 1800 F for 24 hours and a 50 crush strength of at least about 5 pounds.
4. 4 A support according to claim I which possesses a pore volume of about 0 2 to about 0 3 cubic centimeters per gram in pores of from 1000 to 10,000 A in diameter, a pore volume of about 0 6 to about 0 9 centimeters per gram in pores of from 100 to 1000 A in diameter, a pore volume of 0 to about 0 04 cubic centimeters 55 per gram in pores of less than IOOA in diameter, and a total pore volume of about 0.9 to about 1 2 cubic centimeters per gram.
5. A support according to claim 4 which possesses a surface area of about 90 to s, -.b-,-rAl, y- "> -, ' Ä, å v;-s 32 1,603,462 32 about 120 square meters per gram, an attrition loss of less than about 2 %, a volume shrinkage of less than about 4 % upon exposure to a temperature of 18000 F for 24 hours, a compacted bulk density of about 26 to 32 pounds per cubic foot, and crush strength of greater than about 7.
6. 6 A catalyst comprising a support as claimed in any one of claims I to 5 impregnated with at least one catalytically active metal or metal compound.
7. 7 A catalyst according to claim 6 in which the catalytically active metal is platinum, palladium, ruthenium, iridium, rhodium, osmium or a mixture of any two or more thereof.
8. 8 A catalyst according to claim 7 in which the platinum group metal is in an amount up to about 1 O % by weight of the catalyst.
9. 9 A catalyst according to claim 7 or 8 in which the platinum group metal comprises both platinum and palladium.
10. A catalyst according to claim 9 in which the platinum and the palladium are impregnated on separate spheroidal alumina particles.
11. 11 A catalyst according to claim 7 or 8 in which the catalytically active metal comprises platinum, palladium and rhodium.
12. 12 A catalyst according to claim 9 or 11 in which the metals are impregnated such that about 50 % of the total metal surface area is located at depths greater than about 50 microns from the external boundary of the support.
13. 13 A catalyst according to claim 12 in which the metals are impregnated such that about 50 weight percent of the total metals is located at depths greater than about 75 microns from the external boundary of the support.
14. 14 A catalyst according to any one of claims 6 to 13 in which the support is in the form of spheroidal alumina particles as claimed in claim 26 of Application No.
15. 29047.

    A catalyst according to claim 14 in which the support is impregnated with catalytically effective amounts of platinum and palladium such that about 50, of the total metal surface area is located at depths greater than about 50 microns from the external surface of the support.
16. 16 A catalyst according to any one of claims 13 to 15 in which the maximum depth of metal penetration is about 150-400 microns.
17. 17 A catalyst according to claim 16 in which the maximum depth of metal penetration is about 150-250 microns.
18. 18 A catalyst according to any one of claims 13 to 15 in which the maximum depth of penetration of the platinum is about 125-400 microns, the maximum depth of penetration of the palladium is about 125-400 microns, and the maximum depth of penetration of the rhodium is about 125-250 microns.
19. 19 A catalyst according to claim 18 in which the maximum depth of penetration of the platinum is about 125-250 microns, the maximum depth of penetration of the palladium is about 125-250 microns, and the maximum depth of penetration of the rhodium is about 125-200 microns.

    A support according to claim I or a catalyst according to claim 7 substantially as hereinbefore described.

    J A KEMP & CO, Chartered Patent Agents, 14 South Square, Gray's Inn, London, WCIR 5 EU.

    Printed for Her Majesty's Stationery Offlice, by the Courier Press, Leamington Spa 1981 Published by The Patent Office, 25 Southampton Buildings, London, WC 2 A l AY, from which copies may be obtained.

    ,,, f,1 ,-"t;e-- ; 1,603,462